The field of the invention is methods and systems for magnetic resonance imaging. More particularly, the invention relates to methods and systems for parallel magnetic resonance imaging, in which k-space data is acquired from multiple slice locations substantially contemporaneously.
Magnetic resonance imaging (“MRI”) uses the nuclear magnetic resonance (“NMR”) phenomenon to produce images. When a substance such as human tissue is subjected to a uniform magnetic field (polarizing field B0), the individual magnetic moments of the nuclei in the tissue attempt to align with this polarizing field, but precess about it in random order at their characteristic Larmor frequency. If the substance, or tissue, is subjected to a magnetic field (excitation field B1) that is in the x-y plane and that is near the Larmor frequency, the net aligned moment, Mz, may be rotated, or “tipped,” into the x-y plane to produce a net transverse magnetic moment Mxy. A signal is emitted by the excited nuclei or “spins,” after the excitation signal B1 is terminated, and this signal may be received and processed to form an image.
When utilizing these “MR” signals to produce images, magnetic field gradients (Gx, Gy, and Gz) are employed. Typically, the region to be imaged is scanned by a sequence of measurement cycles in which these gradients vary according to the particular localization method being used. The resulting set of received MR signals are digitized and processed to reconstruct the image using one of many well known reconstruction techniques.
The measurement cycle used to acquire each MR signal is performed under the direction of a pulse sequence produced by a pulse sequencer. Clinically available MRI systems store a library of such pulse sequences that can be prescribed to meet the needs of many different clinical applications. Research MRI systems include a library of clinically-proven pulse sequences and they also enable the development of new pulse sequences.
The MR signals acquired with an MRI system are signal samples of the subject of the examination in Fourier space, or what is often referred to in the art as “k-space.” Each MR measurement cycle, or pulse sequence, typically samples a portion of k-space along a sampling trajectory characteristic of that pulse sequence. Most pulse sequences sample k-space in a raster scan-like pattern sometimes referred to as a “spin-warp,” a “Fourier,” a “rectilinear,” or a “Cartesian” scan. The spin-warp scan technique employs a variable amplitude phase encoding magnetic field gradient pulse prior to the acquisition of MR spin-echo signals to phase encode spatial information in the direction of this gradient. In a two-dimensional implementation (“2DFT”), for example, spatial information is encoded in one direction by applying a phase encoding gradient, Gy, along that direction, and then a spin-echo signal is acquired in the presence of a readout magnetic field gradient, Gx, in a direction orthogonal to the phase encoding direction. The readout gradient present during the spin-echo acquisition encodes spatial information in the orthogonal direction. In a typical 2DFT pulse sequence, the magnitude of the phase encoding gradient pulse, Gy, is incremented, ΔGy, in the sequence of measurement cycles, or “views” that are acquired during the scan to produce a set of k-space MR data from which an entire image can be reconstructed.
There are many other k-space sampling patterns used by MRI systems. These include “radial”, or “projection reconstruction” scans in which k-space is sampled as a set of radial sampling trajectories extending from the center of k-space. The pulse sequences for a radial scan are characterized by the lack of a phase encoding gradient and the presence of a readout gradient that changes direction from one pulse sequence view to the next. There are also many k-space sampling methods that are closely related to the radial scan and that sample along a curved k-space sampling trajectory rather than the straight line radial trajectory.
An image is reconstructed from the acquired k-space data by transforming the k-space data set to an image space data set. There are many different methods for performing this task and the method used is often determined by the technique used to acquire the k-space data. With a Cartesian grid of k-space data that results from a 2D or 3D spin-warp acquisition, for example, the most common reconstruction method used is an inverse Fourier transformation (“2DFT” or “3DFT”) along each of the 2 or 3 axes of the data set. With a radial k-space data set and its variations, the most common reconstruction method includes “regridding” the k-space samples to create a Cartesian grid of k-space samples and then performing a 2DFT or 3DFT on the regridded k-space data set. In the alternative, a radial k-space data set can also be transformed to Radon space by performing a 1 DFT of each radial projection view and then transforming the Radon space data set to image space by performing a filtered backprojection.
Depending on the technique used, many MR scans currently require many minutes to acquire the necessary data used to produce medical images. The reduction of this scan time is an important consideration, since reduced scan time increases patient throughout, improves patient comfort, and improves image quality by reducing motion artifacts. Many different strategies have been developed to shorten the scan time.
One such strategy is referred to generally as “parallel MRI” (“pMRI”). Parallel MRI techniques use spatial information from arrays of radio frequency (“RF”) receiver coils to substitute for the spatial encoding that would otherwise have to be obtained in a sequential fashion using RF pulses and magnetic field gradients, such as phase and frequency encoding gradients. Each of the spatially independent receiver coils of the array carries certain spatial information and has a different spatial sensitivity profile. This information is utilized in order to achieve a complete spatial encoding of the received MR signals, for example, by combining the simultaneously acquired data received from each of the separate coils. Parallel MRI techniques allow an undersampling of k-space by reducing the number of acquired phase-encoded k-space sampling lines, while keeping the maximal extent covered in k-space fixed. The combination of the separate MR signals produced by the separate receiver coils enables a reduction of the acquisition time required for an image, in comparison to a conventional k-space data acquisition, by a factor related to the number of the receiver coils. Thus the use of multiple receiver coils acts to multiply imaging speed, without increasing gradient switching rates or RF power.
Two categories of such parallel imaging techniques that have been developed and applied to in vivo imaging are so-called “image space methods” and “k-space methods.” An exemplary image space method is known in the art as sensitivity encoding (“SENSE”), while an exemplary k-space method is known in the art as simultaneous acquisition of spatial harmonics (“SMASH”). With SENSE, the undersampled k-space data is first Fourier transformed to produce an aliased image from each coil, and then the aliased image signals are unfolded by a linear transformation of the superimposed pixel values. With SMASH, the omitted k-space lines are synthesized or reconstructed prior to Fourier transformation, by constructing a weighted combination of neighboring k-space lines acquired by the different receiver coils. SMASH requires that the spatial sensitivity of the coils be determined, and one way to do so is by “autocalibration” that entails the use of variable density k-space sampling.
A more recent advance to SMASH techniques using autocalibration is a technique known as generalized autocalibrating partially parallel acquisitions (“GRAPPA”), as described, for example, in U.S. Pat. No. 6,841,998. With GRAPPA, k-space lines near the center of k-space are sampled at the Nyquist frequency, in comparison to the undersampling employed in the peripheral regions of k-space. These center k-space lines are referred to as the so-called autocalibration signal (“ACS”) lines, which are used to determine the weighting factors that are utilized to synthesize, or reconstruct, the missing k-space lines. In particular, a linear combination of individual coil data is used to create the missing lines of k-space. The coefficients for the combination are determined by fitting the acquired data to the more highly sampled data near the center of k-space.
Nerve tissue in human beings, and other mammals, includes neurons with elongated axonal portions arranged to form neural fibers, or fiber bundles, along which electrochemical signals are transmitted. In the brain, for example, functional areas defined by very high neural densities are typically linked by structurally complex neural networks of axonal fiber bundles. The axonal fiber bundles and other fibrous material are substantially surrounded by other tissue.
Diagnosis of neural diseases, planning for brain surgery, and other neurologically related clinical activities as well as research activities on brain functioning can benefit from non-invasive imaging and tracking of the axonal fibers and fiber bundles. In particular, diffusion tensor imaging (“DTI”) has been shown to provide image contrast that correlates with axonal fiber bundles. In the DTI technique, motion sensitizing magnetic field gradients are applied in a so-called diffusion weighted imaging (“DWI”) pulse sequence so that images, having contrast related to the diffusion of water or other fluid molecules, are obtained. By applying the diffusion gradients in selected directions during the MRI measurement cycle, diffusion weighted images are acquired, from which apparent diffusion tensor coefficients are calculated for each voxel location in the reconstructed images. Since fluid molecules diffuse more readily along the direction of axonal fiber bundles as compared with directions partially or totally orthogonal to the fibers, the directionality and anisotropy of the apparent diffusion coefficients tend to correlate with the direction of the axonal fibers and fiber bundles. Using iterative tracking methods, axonal fibers or fiber bundles can be tracked or segmented using the DTI data.
In order to calculate the apparent diffusion tensor coefficients, however, it is necessary to acquire at least six DWI images using motion-sensitizing gradients directed in six different directions. Indeed, it is desirable to acquire more than six directions, but the acquisition of additional DWI images extends the total scan time beyond what is already a lengthy scan.
In modern diffusion methods such as Q-ball imaging (“QBI”), high angular resolution diffusion imaging (“HARDI”), and diffusion spectrum imaging (“DSI”), an echo planar imaging (“EPI”) pulse sequence is commonly employed for directing the MRI system to acquire diffusion weighted image data. Moreover, the pulse sequence must be repeated up to hundreds of times in order to encode the high angular resolution diffusion information needed, for example, to characterize the fiber bundles in white matter. As a result, these methods are limited by imaging time. For example, a QBI or HARDI acquisition of image data from 64 different slice locations, covering the whole head, and that utilizes 128 different diffusion encoding directions, would require 8,192 repetitions of the EPI sequence. For a standard diffusion protocol, the EPI acquisition for each slice takes about 100 milliseconds (“ms”). Thus, the total acquisition time for the aforementioned imaging session is close to fourteen minutes, which is clinically unacceptable. It would therefore be desirable to provide a method for decreasing the amount of time necessary to acquire the large number of diffusion weighted image data sets demanded by modern diffusion methods.
High spatial resolution functional magnetic resonance imaging (“fMRI”) acquisition methods commonly utilize EPI sequences and, thus, suffer a similar problem. For example, whole-head coverage at high resolution requires as many as 128 slices for 1 mm isotropic image voxels. Since a single slice can be acquired in about 80 ms for fMRI, the temporal resolution for the time series is 128 times 80 ms, which is around 10 seconds. This temporal resolution is often too slow for many functional paradigms, especially event-related paradigms. It would therefore also be desirable to provide a method for increasing the temporal resolution of fMRI acquisitions, including high spatial resolution fMRI acquisitions.
Conventional parallel MRI techniques rely on accelerating standard image acquisitions by undersampling k-space. For example, these methods undersample k-space by reducing the number of phase-encodings acquired during each repetition of a pulse sequence. Data acquisition schemes such as EPI, however, are not sped up significantly by conventional parallel imaging techniques. For example, an EPI sequence is typically only sped up by around 20 percent when using parallel imaging techniques, whereas conventional imaging sequences are sped up by around 200-300 percent. Although k-space can be undersampled in the phase encode direction in EPI acquisitions, the number of slices that can be obtained per second is not significantly increased in this manner. For example, a diffusion acquisition with EPI uses only a relatively small fraction of the acquisition time reading out k-space, while the rest of the time is used for RF excitation and diffusion encoding. Thus, while SENSE and GRAPPA have benefits for reducing image distortion in EPI, they do not significantly speed up a lengthy diffusion acquisition. For this reason, another approach for decreasing the acquisition time required of EPI pulse sequences is desired.
Recently, other methods for decreasing scan time have been developed. For example, methods for the simultaneous acquisition of image data from multiple imaging slice locations, using an array of multiple radio frequency (“RF”) receiver coils, and subsequent separation of the superimposed slices during image reconstruction have be introduced, as described by D. J. Larkman, et al., in “Use of Multicoil Arrays for Separation of Signal from Multiple Slices Simultaneously Excited,” Journal of Magnetic Resonance Imaging, 2001; 13(2):313-317. This method is limited, however, in that the separation of the multiple slices is rendered difficult by the close spatial proximity of the aliased pixels that must be separated during image reconstruction. For example, if image data is acquired from three slices simultaneously, and with an inter-slice spacing of around 3 cm, then aliasing will be present along the slice-encoding direction, and this aliasing must be undone in order to produce reliable images. The origin of the aliased pixels are only 3 cm apart in space, and it is this spatial closeness of the aliased pixels that makes their separation difficult by standard parallel imaging methods, such as sensitivity encoding (“SENSE”).
The failure of SENSE, and other similar methods, to properly separate the aliased pixels results from the differences in detection strength among the multiple array coil elements at the locations of the aliased pixels. In particular, the problem is that the detection profiles of the coil array elements are not unique enough on the spatial scale of a few centimeters. As a result, a high level of noise amplification, characterized by a high SENSE g-factor, is present in the separated images. This result is in contrast to the conventional implementation of SENSE methods, in which an undersampled phase encoding scheme produces aliasing along the phase-encoding direction, which is orthogonal to the slice-encoding direction. Moreover, this in-plane aliasing results in pairs of aliased pixels that are separated by one-half of the image field-of-view (“FOV”). For a conventional brain image, the FOV is equal to around 24 cm; thus, when aliasing occurs in the imaging plane, or slice, the distance between aliased pixels is around 12 cm. It is contemplated that it is the four-fold smaller distance between aliased pairs of pixels that results in significant noise amplification in the method disclosed by Larkman. It would therefore be desirable to provide a method for simultaneous multi-slice imaging that is produces less noise amplification than presently available methods, such as the one taught by Larkman.
Recently, the Larkman method has been improved upon for conventional image acquisition techniques, as described, for example by F. A. Breuer, et al., in “Controlled Aliasing in Parallel Imaging Results in Higher Acceleration (CAIPIRINHA) for Multi-Slice Imaging,” Magn. Reson. Med., 2005; 53(3):684-691. This method, referred to as “CAIPIRINHA,” increases the distance between aliased pixels by introducing a one-half FOV shift in the images of every other slice. This shift is achieved by modulating the phase of the RF excitation pulse used to acquire every other line of k-space by 180 degrees. In this manner, when the image slices superimpose, every-other slice is shifted by one-half of the FOV. Thus, aliased pixels are separated by one-half the FOV in the phase-encoding direction, and are separated, for example, by the 3 cm distance between the slices in slice-encoding direction. This added separation in the phase-encoding direction improves the ability of parallel image reconstruction methods, such as SENSE, to unalias the slices without producing artifacts or significant noise amplification. In addition, the one-half FOV shift also has the benefit that the method does not completely rely on the distribution in coil sensitivity modulation in the slice-encoding direction. As noted above, RF coil array elements typically do not have significant modulation in their sensitivity patterns along, for example, the z-direction, which corresponds to a slice-encoding direction for the commonly acquired axial, or transverse, images.
The CAIPIRINHA method still presents problems, however. For example, the one-half FOV shift imparted to every other slice is achieved by modulating the phase of the RF excitation pulses of every other line in k-space. While this is applicable to conventional acquisitions, in which every line of k-space is acquired with a separate excitation, it is not applicable to EPI acquisitions, in which all the lines of k-space are acquired after a single RF excitation. Thus, the CAIPIRINHA method is not extendable to EPI acquisitions. It would therefore be desirable to provide a method for simultaneous multi-slice imaging that allows a more efficient separation of pixels that are aliased along the slice-encoding direction, but that is applicable to acquisition schemes such as EPI, in which multiple lines of k-space are sampled after each RF excitation pulse.
In addition to the Larkman and CAIPIRINHA methods, there are two other simultaneous multi-slice imaging methods worth noting. One of these methods is the so-called “wideband” method described, for example, by J. B. Weaver in “Simultaneous Multislice Acquisition of MR Images,” Magn. Reson. Med., 1988; 8(3):275-284. The wideband method relies on the application of a constant slice-selective magnetic field gradient during data acquisition. This gradient produces a frequency shift in each imaging slice, thereby shifting the readout bandwidth of any given slice out of the readout bandwidth of the other slices. Applying a slice-selective gradient during readout, however, is problematic. In order to achieve a sufficient frequency shift, the gradient must be strong enough to appreciably tilt the readout gradient direction relative to the phase-encoding and slice-encoding directions. The wideband method has the advantage that parallel imaging methods are not needed since the shift is in the frequency-encoding direction and is sufficient to separate the slices by their frequency. However, the resultant tilting produces trapezoidal pixels, in which the frequency shift is substantial enough to degrade the image resolution. For this latter reason, the wideband method has never found wide usage.
An attempt was made to improve the wideband method by implementing the methodology behind CAIPIRINHA, as described, for example, by R. G. Nunes, et al., in “Simultaneous Slice Excitation and Reconstruction for Single Shot EPI,” Proceedings of the 14th Annual Meeting of ISMRM, Seattle, Wash., USA, 2006, 293. In the Nunes method, a shift of much less than a full FOV by the wideband approach is employed to improve the g-factor penalty in the SENSE based slice separation used by CAIPIRINHA. Nonetheless, a penalty in term of image pixel tilt still exists in the Nunes method. Despite this pixel tilt being smaller than in the wideband method, it is still significant enough to degrade image quality, particularly when the simultaneously excited slices are close in proximity.
The Nunes method adapts the wideband approach by utilizing slice-selective gradients “blips” that are a scaled version of the phase encoded gradients commonly found in EPI acquisitions. Since the slices exist at discrete spatial locations, the slice-selective gradient blips produce a phase ramp in the ky-direction in k-space. This phase ramp is produced across the EPI readout, and is different for each of the excited slices. In this manner, each successively sampled ky line in k-space for a given slice receives a phase increment, Δφ, equal to:Δφ=γGzzt  Eqn. (1);
where γ is the gyromagnetic ratio of the imaged species, Gz is the amplitude of the slice-selective gradient blip, t is the duration of the slice-selective gradient blip, and z is the location of the given slice. Thus, the relative phase increment between even and odd slices is:Δφ=γGzΔzt  Eqn. (2);
where Δz is the inter-slice spacing, or slice separation. Therefore, in order to impart a one-half FOV shift in every other slice requires the relative phase increment between the ky lines to equal 180 degrees.
The problem with this approach of simultaneously applying slice-selective and phase-encoding gradient blips is that this approach results in a rotation of the phase-encoding direction such that it is no longer orthogonal to the slice-selective direction. This rotation results in the same “tilted pixel” problem found in the standard wideband method; however, now the tilting is occurring in the phase-encoding direction, and not the frequency-encoding direction. In this manner, the phase shifts imparted to the ky lines by the slice-selective gradient blips cause through-plane dephasing within each excited slice. While an initial pre-winding gradient can ensure that the amount of through-plane dephasing is zero at the center of k-space, the through-plane dephasing accumulates at the edges of k-space along the ky-direction. As a result, the through-plane dephasing acts as a blurring filter in the phase-encoding direction of the image, similar to how the wideband method produces a blurring filter in the readout direction. As a result of this blurring, the Nunes method utilized only a 15 percent shift in the FOV, instead of the desired one-half FOV shift. Thus, the Nunes method requires a trade-off between the reduction in noise amplification and pixel tilt and blurring. It is noted that even with this limited FOV shift, a significant pixel tilt can result, particularly when more than two slices are excited simultaneously. As a result, the slice achievable separation is limited. It would therefore be desirable to provide a method for simultaneous multi-slice imaging that allows for a greater FOV shift than achievable with the Nunes method without introducing a significant amount of pixel blurring or noise amplification in the reconstructed images.
A second notable method for simultaneous multi-slice imaging was recently described by D. A. Feinberg, et al., in “Simultaneous Echo Refocusing in EPI,” Magn. Reson. Med., 2002; 48(1):1-5. In this method, which is termed “SER-EPI,” the RF excitation of the slices is sequential, as opposed to truly simultaneous. A readout gradient pulse is applied between two sequential excitations, and acts to shift the k-space data of one slice relative to the other along the kx-direction, which corresponds to the readout direction in image space. By lengthening the readout window, the k-space data for both slices is captured sequentially. The data can then be cut apart and reconstructed separately.
This approach has several downsides, however. Since the excitation is not simultaneous, the two slices do not have identical echo times (“TE”). In fact, the TEs typically differ by about 3 ms. This difference in TE is problematic, in that image intensity and contrast is exponentially dependent on TE. Thus, the two slices are not truly identical in image contrast or intensity. Another limitation of the SIR-EPI method is that the lengthened readout needed to capture the shifted k-space data of the second slice increases the total readout duration. In turn, this increased duration increases the B0 susceptibility distortions included in the resultant EPI images. For example, these distortions are increased by a factor of about 1.6 for the simultaneous acquisition of two slices, and by more than a factor of two when three slices are simultaneously acquired. Moreover, for three slice acquisitions, the TE differences are also magnified.
CAIPIRHINA and other simultaneous multi-slice methods have not gained much traction in conventional imaging since there are alternative parallel imaging methods, such as conventional SENSE and GRAPPA, for accelerating standard image acquisitions. However, as noted above, these methods do not confer the same acceleration benefits on pulse sequences such as EPI as they do on other conventional pulse sequences. Unlike parallel imaging methods such as SENSE and GRAPPA, multi-slice acquisition techniques do not aim to shorten the time spent on reading out k-space data, for example, by reducing the number of phase-encodings. Rather, they aim to acquire signal data from multiple image slice locations per acquisition, such that the number of repetitions of a pulse sequence can be reduced to similarly reduce overall scan time. For example, a three-fold accelerated multi-slice acquisition acquires image data from three image slice locations per each repetition of the EPI sequence. As a result of this simultaneous acquisition, the number of repetitions of an EPI sequence required to cover an imaging volume is reduced, thereby similarly reducing the total acquisition time.
It would therefore be desirable to provide a method for simultaneous, multi-slice imaging that significantly decreases the amount of time required for acquiring image data; that is applicable to imaging pulse sequences that sample multiple lines of k-space following each RF excitation, such as EPI sequences; and that allows more reliable separation of aliased pixels than currently available methods for simultaneous multi-slice imaging, so that the benefits associated with these techniques can be realized in a clinical setting.